Clustering is a fundamental task in data science that often arises in real-world applications. In the past decades several alternative formulations of the clustering problem have been proposed. These have led to a variety of algorithms, both exact and approximate, using a broad range of optimization techniques such as local search, core-sets, linear programming, etc. Studying the inherent hardness of clustering has led to deep theoretical results, combining insights from geometry and graph theory.
In this seminar/reading group we will study representative and recent results related to clustering, with a focus on algorithmic and complexity aspects. We present/discuss one or two papers each week.
Requirements: ALP3/HA or similar algorithmic background, mathematical maturity.
Spezialliteratur aus Zeitschriften
T9/SR 005 Übungsraum
wöchentlich, ab 22.10.2021, 14:00 - 16:00 (16 Termine)